JP2005266891A - Region extraction method and device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a region extraction method capable of automatically extracting a desired region accurately by a simple specifying operation such as a one-click operation. <P>SOLUTION: The present invention is an image processing method that extracts a specific region from an image, and has a first region extraction step S12 that extracts an approximate color region from a specified point, a reference point setting step S13 that sets a reference point used for extracting an approximate color region to be added to the extracted approximate color region, and a second region extraction step S14 that extracts an approximate color region from the reference point and adds it to the already extracted approximate color region. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a region extraction method and apparatus suitable for use in extracting a desired region from an image.

  When a user wants to extract a region from an image, a method of specifying a region based on contour information, a method of extracting an approximate color region based on color information, or the like is used. When the contour information is used as a basis, the user can extract a desired region with high accuracy by designating a general region and searching for a contour line adjacent to the general region. However, it is necessary for the user to designate an approximate area, and the operation is complicated especially in the case of two-point designation of a rectangle using a cross key of a portable terminal or the like.

  On the other hand, in area extraction using colors as shown in the example below, for example, the color area differs greatly depending on whether the light source is outdoor sunlight or under a fluorescent lamp in the garden. There was to be.

Many methods have been proposed as an approximate color extraction method. There are ones that use one point in the image as a designated point and obtain an approximate color set based on a set threshold value, and those that are described in Patent Documents 1 to 3, for example. The approximate color region extraction method described in Patent Document 1 is characterized in that the threshold value is set to one by setting the color distance in the color space. The approximate color region extraction method described in Patent Document 2 is characterized in that not only each threshold value of RGB (Red Green Blue) but also a difference value is set and a threshold value is provided. is there. In addition, the method of extracting approximate color regions described in Patent Document 3 pays attention to the fact that the color space recognized as the same color by the human eye differs depending on the position, and depending on the position of the designated point in the color space. The threshold value is changed to a value that allows a human to recognize the same color.
Japanese Patent Laid-Open No. 5-128224 JP-A-11-110552 JP-A-5-236289

  In any of these methods, thresholds equidistant from a specified point are regarded as a set of approximate colors. For this reason, it is possible to successfully extract a face area when the designated point is an intermediate color in an area where bright colors such as faces exist from dark colors to dark colors. When a color close to is designated as a designated point, there is a problem that it cannot be extracted well.

  In addition, in the conventional example, in such a case, it is necessary to specify an area that has not been successfully extracted, and to add an area, such as an input operator for a small information device such as a mobile phone. When an image with limited image designation accuracy is used, there is a problem that a plurality of designation operations and numerical value setting operations become complicated.

  The present invention has been made in view of the above circumstances, and the desired region can be automatically and accurately performed by a simple designation operation such as one click without performing a complicated operation such as setting a threshold value. It is an object of the present invention to provide a region extraction method and apparatus capable of extracting a region.

  In order to solve the above-mentioned problem, the present invention is an image processing method for extracting a specific area from an image, wherein a first area extracting step for extracting an approximate color area from a designated point, and the extracted approximate color area A reference point setting step for setting a reference point used when extracting the approximate color region to be added, and a second region extracting step for extracting the approximate color region from the reference point and adding the approximate color region to the already extracted approximate color region It is characterized by having. According to this, the extraction of the approximate color area is performed separately in the first area extraction step using the designated point and the second area extraction step using the automatically set reference point. This makes it easier to make the approximate color region extraction process more delicate. Therefore, it is possible to extract a region with higher accuracy than in the past by a simple operation in which a user sets a designated point in a desired extraction region.

  Further, the present invention is characterized in that the setting of the reference point in the reference point setting step is performed on the boundary pixels of the region extracted by the first region extraction step. According to this, compared with the case where the reference point is set other than the boundary pixel, it is possible to reduce the overlap of the approximate color regions when the approximate color regions are respectively extracted from the plurality of reference points. Thus, it is possible to reduce the amount of calculation performed in duplicate.

  In the present invention, the setting of the reference point in the reference point setting step is performed on a boundary pixel of the region extracted by the first region extraction step, and a predetermined pixel among the boundary pixels The present invention is characterized in that it is performed for a plurality of pixels apart from each other. According to this, compared with the case where the reference point is set other than the boundary pixel, it is possible to reduce the overlap of the approximate color regions when the approximate color regions are extracted from the plurality of reference points, respectively. Since it is set for each of a plurality of pixels, the amount of calculation can be further reduced.

  In the present invention, the setting of the reference point in the reference point setting step includes a step of obtaining a circumscribed rectangle of the region extracted by the first region extracting step, and dividing the circumscribed rectangle into a mesh pattern in predetermined pixel units. And a step of using as a reference point the intersection of the boundary of the region divided into the mesh and the boundary of the region extracted by the first region extraction step. According to this, since the reference point can be set by obtaining the intersection between the boundary of the approximate color region and the mesh region without tracking the boundary of the approximate color region, for example, the region extracted in the first region extraction step Even if has a complicated boundary shape, the reference point can be set by a relatively simple process.

  In the present invention, the reference point setting in the reference point setting step includes a step of obtaining a circumscribed rectangle of the region extracted by the first region extracting step, a step of obtaining a center of gravity from the circumscribed rectangle, And a step of using, as a reference point, a pixel at a predetermined angle from the center of gravity among the boundary pixels of the region extracted by one region extraction step. According to this, since the reference point can be set by obtaining a pixel at a predetermined angle from the center of gravity in the boundary pixel of the approximate color area without tracking the boundary of the approximate color area, for example, in the first area extraction step Even when the extracted region has a complicated boundary shape, the reference point can be set by a relatively simple process.

  Further, in the present invention, the approximate color region is extracted in the first region extraction step using a predetermined threshold value based on the pixel value of the designated point, and the second region extraction step The approximate color area is extracted using a predetermined threshold value based on the pixel value of the reference point, and the predetermined threshold value in the second area extracting step is the first area. It is smaller than a predetermined threshold value in the extraction step. According to this, since the approximate color area of the first area extraction step extracted from the designated point is relatively larger than each approximate color area extracted in the second area extraction step, Can be expected to improve the accuracy of region extraction.

  Further, the present invention determines whether to end the approximate color region extraction process based on the processing result of the second region extraction step, and determines that the reference point setting step and the first step are not performed when it is determined not to end. The second region extraction step is repeatedly executed. According to this, the reference point setting step and the second region extraction step are repeatedly executed until a predetermined end condition is satisfied. Therefore, it is possible to reduce the case where the approximate color region to be extracted is not extracted. The determination of the end condition can be performed by comparison based on the size of the area of the additional region, the existing edge detection method, the difference in pixel value from the designated point, and the like.

  Further, the present invention is characterized in that the extraction process is determined to end when the area of the approximate color region extracted from the reference point is smaller than a predetermined size. According to this, the end condition can be determined by a simple process of calculating the area of the extraction region (that is, calculating the number of pixels).

  According to the present invention, in the second region extraction step, it is determined whether or not the reference point forms an outline, and if it is determined that the reference point forms an outline, the approximate color region is not extracted from the reference point. It is characterized by that. For example, the edge strength of each reference point is detected by an existing edge detection method, and when the edge strength exceeds a predetermined threshold, it is determined that the reference point forms an outline, and approximate color region extraction processing is performed. Do not. According to this, when the region that the user wants to extract has a clear boundary line, erroneous extraction can be prevented and accurate region extraction can be performed.

  According to another aspect of the present invention, there is provided an image processing apparatus that extracts a specific area from an image, the first area extracting unit that extracts an approximate color area from a specified point, and the extracted approximate color area Reference point setting means for setting a reference point used when extracting the approximate color area to be added, and second area extraction means for extracting the approximate color area from the reference point and adding it to the already extracted approximate color area. It is characterized by that.

  According to another aspect of the present invention, there is provided a program for executing an image processing method for extracting a specific area from an image, the first area extracting step for extracting an approximate color area from a specified point, A reference point setting step for setting a reference point used when extracting an approximate color area to be added to the approximate color area; and a second that extracts an approximate color area from the reference point and adds the approximate color area to the already extracted approximate color area It includes a description for executing the region extraction step using a computer.

  Hereinafter, an embodiment of a region extraction device of the present invention will be described with reference to the drawings. The region extraction device of the present invention is realized by a computer constituting an information device such as a mobile phone, a portable information terminal, a digital camera, and a personal computer, its input device, and a program executed by the computer. More specifically, it is implemented as one image processing function in each device.

  FIG. 1 is a block diagram showing the configuration of the region extraction apparatus of the present embodiment. Each block is composed of a combination of hardware and software. In the present embodiment, a process of extracting a specific area from an image designated in accordance with a user operation is executed by a combination of these blocks.

The image input unit 1 is configured to perform processing for inputting or specifying an image to be processed.
The image input unit 1 has a function of inputting and selecting or selecting an image to be processed using an operation key or a display device mounted on an information device realizing the present embodiment and a storage device or a detachable recording medium. I will provide a. The designated point setting unit 2 displays an image input by the image input unit 1 and selected by the user as a processing target on the display device, and displays one point (one pixel or a plurality of pixels) in the display image in accordance with a user operation. 1 area) is designated.

  The first approximate color area extraction unit 3 extracts an approximate color area composed of a plurality of continuous pixels having the specified point and the approximate color with reference to the specified point set by the specified point setting unit 2. Perform the process. The extraction of the approximate color area by the first approximate color area extraction unit 3 determines whether the target pixel has a pixel value within a predetermined threshold range with reference to the pixel value of the designated point. While done.

  The reference point setting unit 4 is newly added in the second approximate color region extraction unit 5 to the approximate color region extracted by the first approximate color region extraction unit 3 or the second approximate color region extraction unit 5. It has a function of setting a reference point used when extracting the approximate color area. One or a plurality of pixels in the image input by the image input unit 1 are set as a reference point. The reference point setting unit 4 first sets a reference point based on the approximate color region extracted with the designated point as a reference by the first approximate color region extraction unit 3. The reference point is set, for example, by setting all of the extracted boundary pixels in the approximate color region or by using a plurality of pixels that are separated from the boundary pixels by a predetermined pixel as the reference point.

  The second approximate color area extraction unit 5 extracts one or a plurality of approximate color areas from one or a plurality of reference points set by the reference point setting unit 4 and adds them to the already extracted approximate color areas. A function of determining an approximate color area; The second approximate color area extraction unit 5 extracts an approximate color area composed of a plurality of continuous pixels having the reference point and the approximate color and using the reference point set by the reference point setting unit 4 as a reference. Perform the process. The extraction of the approximate color region by the second approximate color region extraction unit 5 determines whether or not the pixel to be determined has a pixel value within a predetermined threshold range with reference to the pixel value of the reference point. While done. At this time, the threshold value used in the second approximate color region extraction unit 5 is set to be smaller than the threshold value used in the first approximate color region extraction unit 3.

  The end determination unit 6 determines whether or not to end the approximate color region extraction process based on the processing result of the second approximate color region extraction unit 5, and determines that the end does not end with the reference point setting unit 4 The second approximate color area extraction unit 5 has a function of controlling to repeatedly execute the process. The end determination unit 6 determines whether or not to end, for example, the area (that is, the number of pixels) of the approximate color region extracted from each reference point in the second approximate color region extraction unit 5 is smaller than a predetermined size. In this case, it is determined that the condition to end is satisfied. However, the determination method can also be performed by comparison with an existing edge detection method or a predetermined value based on a pixel value difference from a designated point.

  The extraction area output unit 7 determines the approximate color area extracted by the second approximate color area extraction unit 5 based on the designated point designated by the user when the end determination unit 6 determines to end the extraction process. Output as the final approximate color area extracted. For example, the area is displayed on the display device of the device of the present embodiment, or the process of handing over data indicating the area to another program that performs image processing on the input image based on the extracted area is performed.

  Next, an operation example of the embodiment shown in FIG. 1 will be described with reference to the flowcharts shown in FIGS. In this example, a user first operates an information device to call an image input function by the image input unit 1 of the region extraction apparatus, inputs (or selects) an image to be processed, and then displays the image on the screen of the display device. Display an image. Then, an operator such as a cross cursor is operated to move the cursor to an area to be extracted by the user in the image and click the operator. Then, one point in the image is designated by the designated point setting unit 2 in FIG. 1 (step S11). FIG. 10 shows an example of the image 10 to be processed. In the example of FIG. 10, it is assumed that the user is trying to extract the extraction target area 11 in the image 10. It is assumed that the designated point 12 is set according to the user's operation.

  Next, the first approximate color area extraction unit 3 automatically performs a process of extracting an approximate color area from a designated point 12 (for example, RGB values, R: 235, G: 176, B: 170) ( Step S12). For example, the approximate color region is extracted by setting the threshold value to “32” or the like and the color space to be RGB, Lab, or the like. However, in this example, the RGB color space is used. The Lab color space is a method for expressing a color by two color elements, black to white (lightness), an element (a) in the range from green to red, and an element (b) in the range from blue to yellow. .

  For example, if the pixel at the designated point 12 has R: 235, G: 176, and B: 170, and the threshold value is 32, the pixel values are R: 203 to 267, G: 144 to 208, B A pixel in the range of 138 to 202 and continuous with the designated point 12 is extracted as an approximate color region. In the example of FIG. 4, it is assumed that the approximate color region 13 shown by shading is extracted.

  In the above case, for example, pixels having pixel values R: 202, G: 176, and B: 170 are approximate color ranges (R: 203 to 267, G: 144 to 208) determined by the threshold value “32” from the designated point 12. , B: 138-202), so it is not extracted. However, the designated point 12 is arbitrarily set from a plurality of pixels by the user's operation, and when the threshold value is set to a predetermined fixed value, the desired extraction target region 11 has a slight amount like this. It is conceivable that many pixels having pixel values outside the approximate color range are included. For example, when an approximate region is to be extracted in a single process, it is necessary to change the threshold value to, for example, “33” in order to extract the region including such pixels. However, it is necessary to set an appropriate threshold value for each desired region, which is a troublesome operation. Therefore, in the present embodiment, the following processing is automatically performed so that even such a pixel can be extracted without changing the threshold value.

  As described above, after the approximate color region extraction unit 3 performs the process of extracting the approximate color region 13 from the designated point 12 (step S12), the reference point setting unit 4 performs the reference point 15 ( The process for setting (FIG. 4) is performed (step S13).

  In order to set the reference point 15, first, the boundary pixel of the extraction region (approximate color region 13) is marked. That is, the boundary pixel 14 that is the boundary of the extracted approximate color region 13 shown in FIG. 4 is extracted. The determination of whether or not the pixel is the boundary pixel 14 includes, for example, a pixel 16 that is not the extracted approximate color region 13 in four adjacent pixels (or eight pixels) as shown in FIG. 5A or 5B. Depending on whether or not the pixel 16 is included, the boundary pixel 14 is marked.

  The example of FIG. 5A shows a case where the boundary pixel 14 is determined with reference to four adjacent pixels. For example, regarding the pixel 17a in the approximate color region 13, the pixel 17a is the boundary pixel 14 because the pixel 16 outside the region is included in the range 18a including the pixel 17a and four pixels of the upper, lower, left, and right pixels. Determined. On the other hand, for the pixel 17b in the approximate color area 13, the pixel 17b is not the boundary pixel 14 because the pixel 16 outside the area is not included in the range 18b including the pixel 17b and the four pixels of the upper, lower, left, and right pixels. Determined.

  On the other hand, the example of FIG. 5B shows a case where the boundary pixel 14 is determined based on the adjacent eight pixels. For example, for the pixel 17a in the approximate color area 13, since the pixel 16 outside the area is included in the range 19a that includes the pixel 17a and eight pixels of each of the upper, lower, left, and right and diagonal upper, lower, left, and right pixels, the pixel 17a is a boundary pixel. 14 is determined. For the pixel 17b in the approximate color area 13, the pixel 17b is also a boundary pixel because the pixel 16 outside the area is included in the range 19b that includes the pixel 17b and eight pixels of each of the upper, lower, left, and right and diagonal up, down, left, and right pixels. 14 is determined.

  The reference point 15 may be all of the extracted boundary pixels 14 as reference points, or may be a limited number of pixels selected according to a predetermined condition. In the example shown in FIGS. 5A and 5B, the reference point 15 is set for every five pixels of the boundary pixel 14.

  Next, the second approximate color area extraction unit 5 performs a process of extracting the second approximate color area from the reference point 15 (step S14). Here, with reference to FIG. 3, the process of extracting the second approximate color area in step S14 of FIG. 2 will be described.

  In the process of extracting the second approximate color area (step S14 in FIG. 2), the approximate color area is sequentially extracted from one or a plurality of reference points set in step S12 in FIG. The process of sequentially adding to the approximate color area is performed. Since the reference point set in step S13 is automatically set based on a certain shape condition without considering the pixel value, for example, the approximate color region extracted at a certain reference point and other reference points The approximate color region extracted from the point may overlap. In this embodiment, in such a case, with respect to a reference point included in an approximate color area from another reference point, the process of extracting the approximate color area from the reference point is omitted. Therefore, first, before the approximate color region extraction process, it is determined whether or not the reference point is not included in the approximate color regions from other reference points (step S21). Next, only when the reference point is not included in the approximate color area from another reference point (that is, only when the reference point is at the boundary of the approximate color area) ("is a boundary" in step S21), FIG. Similar to step S12 (however, using a reference point), an approximate color region extraction process is performed (steps S22 to S24). And the process which adds the extracted approximate color area | region to the already extracted approximate color area | region is performed (step S25). These processes are repeatedly executed for all the reference points (“Yes” in step S26 to step S21 and subsequent steps (or “No” in step S26)).

  FIG. 6 is an explanatory diagram for explaining the process of extracting the second approximate color region. In FIG. 6, the same reference numerals are used for the same components as those in FIGS. 4 and 5A and 5B. In this example, the second reference color region 21a is extracted at the reference point 15a by using the first reference point as the pixel located at the top of the image (the leftmost of the pixels when there are multiple images). Yes. The pixel value of the reference point 15a is acquired (step S22), and a color range is set from the acquired pixel value (step S23). At this time, the color range is set by using a threshold (such as “16”) smaller than the first region extraction step (step S12 in FIG. 1). Next, the approximate color region 21a continuous from the reference point 15a in the set color range is extracted (step S24). Then, the extracted approximate color area 21a is added to the already extracted approximate color area (in this case, the approximate color area 13 because it is the first reference point) (step S25).

  In the present embodiment, if a reference point is selected clockwise from the first reference point, the reference point to be processed next to the reference point 15a is the reference point 15b. However, since the reference point 15b is included in the approximate color region 21a from the reference point 15a, it is determined that it is not a boundary in the boundary determination (step S21), and therefore the approximate color region is not extracted. Next, since the reference point 15c is a boundary (because it is not included in other approximate color regions) (“is a boundary” in step S21), the approximate color region 21b is extracted from the reference point 15c (steps S22 to S24). The approximate color region 21b is added to the already extracted approximate color region (in this case, the region where the approximate color region 21a is added to the approximate color region 13) (step S25).

  Thereafter, similarly, the process of extracting the approximate color area from the reference points 15d, 15e,... Is performed, and the extracted approximate color area 21c and the like are added to the already extracted approximate color area. When the processes for all the reference points are completed (“None” in step S26), the second approximate color area extraction process is completed. FIG. 6 shows a state in which the processing for all the reference points has been completed.

  Next, the end determination unit 6 performs processing to determine whether or not to end the approximate color region extraction process based on the processing result by the second approximate color region extraction unit 5 in step S14 of FIG. Step S15). If it is determined to end, the result is output and the process ends (“END” in step S15). On the other hand, if it is determined not to end (“repeat” in step S15), the reference point setting unit 4 resets the reference point, and then the second approximate color region extraction unit 5 uses the new reference point. The approximate color area is repeatedly extracted and further added to the approximate color area (steps 13 to 13). The end determination unit 6 determines whether or not to end, for example, from each reference point (reference points 15a, 15c, 15e, etc. in FIG. 4) extracted in the second approximate color region extraction process in step S14. When the area (namely, the number of pixels) of the extracted approximate color regions (approximate color regions 21a, 21b, 21c, etc.) is smaller than a predetermined size, the determination is made that the condition for termination is satisfied. The area is determined for each of the approximate color regions 21a, 21b, 21c,... So that the end condition is satisfied when all the areas are smaller than a predetermined value, It can be determined that the end condition is satisfied when the total area of 21b, 21c,... Is smaller than a predetermined value. The predetermined value of the area can be set by the ratio (1% or the like) to the already extracted approximate color region in addition to the number of pixels (10 pixels or the like).

  Note that the end determination method may be performed by determining whether or not each reference point used in the second region extraction step S14 is near the contour of the desired region. For example, the contour value is calculated at each reference point (Sobel, Labrador, etc.), and if each contour value is equal to or greater than a predetermined threshold value, the end condition can be satisfied.

  Further, in the second region extraction step of step S14, when the approximate color region region extraction process from the reference point is performed, pixels having pixel values that are separated from the pixel value of the specified point by a predetermined threshold or more are approximated. Processing to be excluded from the color area can also be added.

  In the present embodiment, the pixel value of the reference point is, for the above example, just the approximate color range such as R: 203, G: 176, B: 170, and R: 35, G: 25, B. Can be roughly divided into pixels in the vicinity of an outline having a completely different pixel such as 75. The pixel near the contour may be near the center of the initial approximate color region (approximate color region 13). In the present embodiment, when the pixel near the contour becomes the reference point, the threshold value is set to a value smaller than that used in the first region extraction (“32”) such as “16”. , R: 35, G: 25, B: 75 and the like are not newly added.

  On the other hand, in the case where the pixel value of the reference point is a pixel near the approximate color range from the designated point 12 such as R: 203, G: 176, B: 170, the adjacent pixels of the pixel are R: 202, G: 176, There are many cases where the color changes smoothly, such as B: 170. Although such a part is considered as the same area for human beings, it is an effect of the present embodiment that such an area can be automatically extracted and added. In this case, since the boundary pixels are R: 203, G: 176, and B: 170, if the second approximate color range based on the reference point is the threshold value 16, R: 187 to 219 , G: 160 to 192, and B: 154 to 186. Thus, since a slightly different approximate color region can be extracted from the first approximate color range (R: 203 to 267, G: 144 to 206, B: 138 to 202) based on the designated point, It is possible to extract a desired region with high accuracy.

  Next, another configuration example of the reference point setting unit 2 in FIG. 1 will be described. In this configuration example, the reference point setting in the reference point setting step S13 includes a step of obtaining a circumscribed rectangle of the approximate color region extracted in the first region extracting step S12, a step of obtaining a center of gravity from the circumscribed rectangle, And a step of setting a pixel at a predetermined angle from the center of gravity among the boundary pixels of the region extracted in one region extraction step S12 as a reference point. With reference to FIG. 7, an example of setting a reference point when this configuration is used will be described. In the example illustrated in FIG. 7, eight pixels at every 45 ° are set as reference points 15 a to 15 h from the center of gravity 32 of the circumscribed rectangle 31 obtained for the approximate color region 13. In FIG. 7, the same components as those shown in FIG. This configuration example is advantageous in terms of processing speed because it is not necessary to track boundary pixels as compared with the above configuration described with reference to FIGS. 5 (a) and 5 (b).

  Next, still another configuration example of the reference point setting unit 2 in FIG. 1 will be described. In this configuration example, the setting of the reference point in the reference point setting step S13 is a step of obtaining a circumscribed rectangle of the approximate color area extracted in the first area extracting step S12, and the circumscribed rectangle is reticulated in predetermined pixel units. The dividing step includes a step of using the intersection of the boundary of the region divided into the mesh and the boundary of the region extracted in the first region extraction step S12 as a reference point. With reference to FIG. 8, a setting example of a reference point when this configuration is used will be described. In the example shown in FIG. 8, the circumscribed rectangle 31 obtained for the approximate color region 13 is divided into a mesh pattern with a predetermined pixel length, and the intersection points of the sides of the divided shape and the boundary pixels 14 are the reference points 15a, 15b,. It is said. In FIG. 8, the same reference numerals are used for the same components as those shown in FIG. In this configuration example, it is not necessary to calculate the angle as compared with the above configuration described with reference to FIG.

  The embodiment of the present invention is not limited to the above-described embodiment, and can be changed by dividing or integrating each block or each step. The program constituting the present invention can be distributed via a computer-readable recording medium or a communication line.

The block diagram which shows the structure of embodiment of the area | region extraction apparatus by this invention. The flowchart for demonstrating the operation example of embodiment of FIG. The flowchart for demonstrating the content of step S14 of FIG. Explanatory drawing which shows the image 10 for demonstrating the operation example of a structure of FIG. Explanatory drawing for demonstrating the example of a setting of the reference point 15 of FIG. Explanatory drawing which shows the image 10 for demonstrating the processing content shown in FIG. Explanatory drawing which shows the image 10 for demonstrating the other structure of the reference point setting part 4 of FIG. Explanatory drawing which shows the image 10 for demonstrating another structure of the reference point setting part 4 of FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Image input part, 2 Specified point setting part, 3 1st approximate color area extraction part, 4 Reference point setting part, 5 2nd approximate color area extraction part, 6 End determination part, 7 Extraction area output part, 12 designation | designated Point, 15, 15a-15h Reference point

Claims (11)

An image processing method for extracting a specific area from an image,
A first region extraction step for extracting an approximate color region from a designated point;
A reference point setting step for setting a reference point used when extracting an approximate color area to be added to the extracted approximate color area;
A region extraction method comprising: a second region extraction step of extracting an approximate color region from the reference point and adding the approximate color region to an already extracted approximate color region.   2. The region extraction method according to claim 1, wherein the reference point setting in the reference point setting step is performed for a boundary pixel of the region extracted by the first region extraction step.   The setting of the reference point in the reference point setting step is performed on the boundary pixels of the region extracted by the first region extraction step, and a plurality of pixels separated by a predetermined pixel among the boundary pixels are set. The region extraction method according to claim 1, wherein the region extraction method is performed for the region. The setting of the reference point in the reference point setting step includes a step of obtaining a circumscribed rectangle of the region extracted by the first region extracting step, a step of dividing the circumscribed rectangle into a mesh shape in units of predetermined pixels, and the mesh shape The region extraction method according to claim 1, further comprising a step of using an intersection of the boundary of the region divided into two and the boundary of the region extracted by the first region extraction step as a reference point. The setting of the reference point in the reference point setting step includes the steps of obtaining a circumscribed rectangle of the region extracted by the first region extracting step, obtaining a centroid from the circumscribed rectangle, and the first region extracting step. The region extraction method according to claim 1, further comprising the step of using a pixel at a predetermined angle from the center of gravity of the boundary pixels of the extracted region as a reference point. The approximate color region extraction in the first region extraction step is performed using a predetermined threshold value based on the pixel value of the designated point;
Extraction of the approximate color area in the second area extraction step is performed using a predetermined threshold value based on the pixel value of the reference point;
The predetermined threshold value in the second region extraction step is smaller than the predetermined threshold value in the first region extraction step. 6. Region extraction method. Based on the processing result of the second region extraction step, it is determined whether or not the approximate color region extraction processing is to be ended, and when it is determined not to end, the reference point setting step and the second region extraction step are The region extraction method according to claim 1, wherein the region extraction method is repeatedly executed.   The region extraction method according to claim 7, wherein the extraction process is determined to be terminated when the area of the approximate color region extracted from the reference point is smaller than a predetermined size. In the second region extraction step, it is determined whether or not the reference point forms an outline, and when it is determined that the reference point forms an outline, the approximate color region is not extracted from the reference point. Item 9. The region extraction method according to any one of Items 1 to 8. An image processing apparatus for extracting a specific area from an image,
First area extracting means for extracting an approximate color area from a designated point;
A reference point setting means for setting a reference point used when extracting an approximate color region to be added to the extracted approximate color region;
A region extracting apparatus comprising: a second region extracting unit that extracts an approximate color region from the reference point and adds the approximate color region to the already extracted approximate color region. A program for executing an image processing method for extracting a specific area from an image,
A first region extraction step for extracting an approximate color region from a designated point;
A reference point setting step for setting a reference point used when extracting an approximate color area to be added to the extracted approximate color area;
A region extraction program comprising: a description for executing, using a computer, a second region extraction step of extracting an approximate color region from the reference point and adding the approximate color region to an already extracted approximate color region.

Images